Endothelial cell–restricted disruption of FoxM1 impairs endothelial repair following LPS-induced vascular injury

Recovery of endothelial integrity after vascular injury is vital for endothelial barrier function and vascular homeostasis. However, little is known about the molecular mechanisms of endothelial barrier repair following injury. To investigate the functional role of forkhead box M1 (FoxM1) in the mechanism of endothelial repair, we generated endothelial cell–restricted FoxM1-deficient mice (FoxM1 CKO mice). These mutant mice were viable and exhibited no overt phenotype. However, in response to the inflammatory mediator LPS, FoxM1 CKO mice displayed significantly protracted increase in lung vascular permeability and markedly increased mortality. Following LPS-induced vascular injury, FoxM1 CKO lungs demonstrated impaired cell proliferation in association with sustained expression of p27^Kip1 and decreased expression of cyclin B1 and Cdc25C. Endothelial cells isolated from FoxM1 CKO lungs failed to proliferate, and siRNA-mediated suppression of FoxM1 expression in human endothelial cells resulted in defective cell cycle progression. Deletion of FoxM1 in endothelial cells induced decreased expression of cyclins, Cdc2, and Cdc25C, increased p27^Kip1 expression, and decreased Cdk activities. Thus, FoxM1 plays a critical role in the mechanism of the restoration of endothelial barrier function following vascular injury. These data suggest that impairment in FoxM1 activation may be an important determinant of the persistent vascular barrier leakiness and edema formation associated with inflammatory diseases.

The endothelial monolayer at the interface between extravascular space and blood plays a crucial role in vascular homeostasis (1). The endothelial monolayer helps to maintain the antithrombotic and antiinflammatory state of the microvascular bed, and it additionally controls the tone and proliferative state of the underlying vascular smooth muscle cells. In pathologic situations, such as the initiation and progression of atherosclerosis and acute lung injury (ALI), vascular endothelial injury and increased permeability of the barrier promote the extravasation of neutrophils and monocytes across the vessel wall (2, 3). This loss of endothelial barrier function is associated with severe consequences: tissue edema formation, thrombus formation, neointimal thickening, and abnormal responses to endothelium-dependent agonists (4–6). Intractable endothelial injury characterized by persistently increased lung vascular permeability is a hallmark of ALI (7). Reversibility of endothelial injury reflects the severity of injury itself as well as the capacity of cells to undergo the repair necessary for reannealing the endothelial barrier. However, the molecular mechanisms that regulate endothelial cell proliferation and barrier repair following vascular injury are poorly understood.

Forkhead box M1 (FoxM1 or FoxM1B; previously known as HFH-11/Trident/Win/MPP2) is a member of the mammalian fox family of transcription factors that share homology in the winged helix or forkhead DNA–binding domain (8, 9). FoxM1 is expressed during cellular proliferation and is silenced in terminally differentiated cells (10–12). During liver regeneration, FoxM1 is essential for hepatocyte DNA replication and entry into mitosis (13). Mouse hepatocytes deficient in FoxM1 failed to develop hepatocellular carcinomas in response to carcinogens as a result of impaired hepatocyte proliferation (14). Another study using a transgenic mouse model demonstrated that overexpression of FoxM1 accelerates the proliferation of distinct pulmonary cell types following butylated hydroxytoluene–induced lung injury (15). Previous studies have not addressed whether FoxM1 has a role in the mechanism of repair of the endothelial barrier following vascular injury, as the FoxM1-null mutation causes embryonic lethality due to impaired execution of mitosis (16). Studies demonstrate that FoxM1 is essential for the expression of cell division cycle 25A (Cdc25A) phosphatase and diminishment of the nuclear accumulation of the cyclin-dependent kinase inhibitor proteins p21^Cip1 and p27^Kip1 required for progression into DNA replication (13, 14). During the G2/M transition, FoxM1 regulates transcription of Cdc25B phosphatase responsible for stimulating cyclin-dependent kinase 1 (Cdk1) activity and inducing expression of aurora B kinase, survivin, and polo-like kinase 1, which in turn phosphorylates substrates essential for mitosis (13, 14, 17–19). Herein, using mice with endothelial cell–restricted FoxM1 deficiency (FoxM1 CKO mice), we addressed the possibility that FoxM1 regulates endothelial cell proliferation after injury and thereby promotes the recovery of lung microvasculature endothelial barrier function. FoxM1 CKO mice were generated using Tie2 promoter/enhancer-driven Cre transgenic mice. Following LPS-induced lung microvascular injury, FoxM1 CKO mice exhibited markedly impaired endothelial barrier recovery due to defective endothelial cell proliferation associated with increased expression of p27^Kip1 and decreased expression of cyclin B1 and Cdc25C. FoxM1 CKO mice also showed increased mortality in response to LPS challenge in association with severe pulmonary edema. These data demonstrate for what we believe is the first time that FoxM1 is essential in regulating the proliferation of endothelial cells subsequent to an inflammatory insult and that it thereby plays an important role in reestablishing endothelial barrier integrity following vascular injury. The results point to the value of therapeutic strategies targeting FoxM1 to promote endothelial barrier reannealing and resolve nflammation.

Endothelial cell–specific disruption of FoxM1. To gain insights into the function of FoxM1 in endothelial repair following vascular injury, we inactivated FoxM1 in the mouse endothelium. Mice carrying a FoxM1 gene in which exons 4–7 were flanked by 2 loxP sites (13) were bred with Tie2 promoter/enhancer-driven Cre transgenic mice (20). Tie2 promoter/enhancer-driven Cre expression induced endothelial cell–specific deletion of mouse FoxM1 loxP/loxP-targeted allele by removing the entire winged-helix DNA–binding and C terminal transcriptional activation domains (13). As shown in Figure 1A, the predicted recombination-derived 7.6-kb band was identified in Southern blots only in endothelial cell–rich tissues such as aorta and lung from FoxM1 CKO. The WT 2.9-kb band was more prominent in cells that did not express Cre. Quantitative real-time RT-PCR (QRT-PCR) revealed a 50% reduction in FoxM1 mRNA expression in FoxM1 CKO lungs whereas expression of FoxO1, a ubiquitously expressed Fox, was unaffected (Figure 1B). As shown in Figure 1C, we observed 70% reduction of FoxM1 mRNA expression in primary cultures of endothelial cells isolated from FoxM1 CKO lungs but not in nonendothelial cells (fibroblasts and epithelial cells). As the primary culture of FoxM1 CKO endothelial cells was assessed as 90% pure, the efficiency of FoxM1 deletion was greater than 90%, a level comparable to that given in other reports of Tie2-Cre endothelial-specific gene deletions (20–22).

Figure 1

Mouse model of endothelial cell–restricted deletion of FoxM1. (A) Southern blot demonstrating recombination of the FoxM1 floxed allele in endothelial cell–enriched tissues. Mouse genomic DNA (15 μg /lane) was digested with Bgl II and XbaI. The blots were probed with 3′ probe as described (13). Tie2 promoter/enhancer-driven Cre expression created a predicted 7.6-kb band in the aorta and lungs. The pseudogene band is indicated by an asterisk. (13). fl, FoxM1 floxed allele; AA, aorta; L, lung. (B) Quantitative analysis of mRNA expression of FoxM1 and FoxO1 by QRT-PCR. RNA was extracted from lungs of either FoxM1 WT or FoxM1 CKO mice. mRNA expression of FoxM1 and FoxO1 was normalized to mouse cyclophilin mRNA. There were no differences in FoxM1 mRNA levels in mice with the 3 genotypes FoxM1^+/+, FoxM1^fl/+, and FoxM1^fl/fl; therefore, all these mice were referred to as WT mice. Data are expressed as mean ± SD, n = 3. **P < 0.05. CKO, FoxM1 CKO. (C) Quantitative analysis of FoxM1 expression in population of cells either enriched or depleted of endothelial cells. RNA was extracted from an endothelial cell–enriched primary culture (EC, passage 0) isolated from mouse lungs or from nonendothelial cell primary cultures (non-EC, passage 4), fibroblasts, and epithelial cells. Approximately 70% reduction of FoxM1 mRNA levels was detected in FoxM1 CKO EC and not in FoxM1 CKO non-EC (n = 3), indicating endothelial cell–specific knockdown of FoxM1. ^#P < 0.01. n = 3.

FoxM1 CKO mice have normal cardiovascular systems and basal lung vascular permeability. Unlike the FoxM1-null mutation mice (16, 23), approximately 80% of FoxM1 CKO were viable and indistinguishable from their WT littermates. FoxM1 CKO mice grew to adulthood, living as long as WT mice (≥ 18 months). FoxM1 CKO hearts and lungs from adult mice (4–5 months) were histologically normal (Figure 2E and data not shown); heart/body weight (BW) ratio was not different from that of WT (data not shown). To evaluate endothelial barrier function, we quantified microvascular permeability in the murine isolated perfused lung model (24, 25). The basal lung capillary filtration coefficient (K_f,c) of FoxM1 CKO was similar to that of WT (Figure 2A).

Figure 2

Impaired lung vascular endothelial barrier reannealing following LPS-induced microvascular injury in FoxM1 CKO mice. (A) Time course of increase in lung microvessel permeability as measured by K_f,c. Basal, 0.2 ml PBS; LPS, 5 mg/kg BW. Data are expressed as mean ± SD, n = 3–5 per group. *P < 0.001 versus basal; **P < 0.001, CKO versus WT; ^#P < 0.05, CKO versus WT. (B) Representative graphs of mouse lung wet weights continuously recorded during measurement of microvessel permeability. 48h-LPS, 48 hours after LPS challenge. (C and D) Time course of loss of isogravimetric state of lungs (edema formation) obtained from mouse lungs at indicated times after LPS challenge. WT (C) and FoxM1 CKO mice (D) are represented. Points indicate mean ± 1 SD, n = 3–5 mice per group. ^#P < 0.05 WT or CKO 6 hours after LPS challenge (6h-LPS) versus basal; ^##P < 0.05 CKO 48 hours after LPS challenge (48h-LPS) versus CKO basal or WT 48 hours after LPS challenge; ^ΧP < 0.05 CKO 96 hours after LPS challenge (96h-LPS) versus CKO basal or WT 96 hours after LPS challenge. (E) Representative micrographs of H&E staining show perivascular leukocyte infiltration in FoxM1 CKO lungs 96 hours after LPS challenge (5 mg/kg). Arrows indicate leukocyte infiltration. Br, bronchia; S, small vessels (<150 μm in diameter); L, large vessels (>150 μm in diameter). Scale bar: 50 mm. (F) Quantitative analysis of infiltrating leukocytes in lungs at 96 hours after LPS challenge. Bar graphs show infiltrating leukocytes in vessels of different diameters. Percentages of vessels of different diameters exhibiting leukocyte infiltration are shown. Data are expressed as mean ± SD, n = 3–4. ^#P < 0.05 versus WT (infiltrating leukocytes per vessel); ^##P < 0.05 versus WT (% of vessels).

FoxM1 is required for endothelial barrier reannealing following LPS-induced vascular injury. To determine whether FoxM1 was induced after the onset of vascular injury, we challenged mice by i.p. injection with LPS, the gram-negative bacterial cell wall component, to induce ALI (24, 26, 27). QRT-PCR revealed a 4-fold increase of FoxM1 mRNA expression in WT lungs 3 days after LPS challenge at a dose of 5 mg/kg BW (Figure 3A). The induction was not detected until day 2 and peaked at day 3 after LPS challenge (Figure 3A and Supplemental Figure 1; supplemental material available online with this article; doi:10.1172/JCI27154DS1). Recruitment of inflammatory cells into lungs and vascular leakage are known to occur as early as 3 hours after LPS challenge with maximal lung microvascular injury observed at approximately 24 hours (24, 28, 29). Thus, the time course of FoxM1 expression (Figure 3A) paralleled the recovery phase of LPS-induced injury. In contrast to WT lungs, there was a significantly lower FoxM1 expression detected at day 3 and day 4 after LPS challenge in FoxM1 CKO lungs (Figure 3A), reflecting the presence of FoxM1 expression in the nonendothelial cell population. In addition, 2 other inflammatory mediators, TNF-α and H₂O₂, significantly induced FoxM1 mRNA in primary cultures of endothelial cells isolated from WT mouse lungs (Figure 3B).

Figure 3

Time course of FoxM1 expression following LPS-induced lung microvascular injury. (A) Quantitative analysis of FoxM1 mRNA levels in mouse lungs by QRT-PCR. Total RNA was extracted from lungs collected from either WT or FoxM1 CKO mice after LPS exposure (5 mg/kg BW) at indicated times. FoxM1 mRNA levels were normalized to cyclophilin. Data are expressed as mean ± SD, n = 3–4. *P < 0.05 WT versus CKO; **P < 0.05 versus basal WT. (B) FoxM1 mRNA expression in endothelial cell–enriched primary cultures isolated from WT mouse lungs. Total RNA was extracted from endothelial cell–enriched primary cultures (passage 2) treated with either PBS (basal), 1,000 U/ml of recombinant mouse TNF-α, or 10 μM of H₂O₂ for 22 hours. Data are expressed as mean ± SD, n = 3. *P < 0.05.

As seen in the acute phase of ALI in humans (7), endothelial injury induced by LPS is characterized by endothelial cell apoptosis and increased endothelial permeability (24, 28–30). We therefore assessed these responses in FoxM1 CKO lungs. After LPS challenge, there was no significant difference in the level of apoptosis in lung sections of FoxM1 CKO and WT mice at either day 1 (the peak of apoptosis response in both groups; Figure 4, A and B) or at later time points (data not shown). As shown in Figure 4C, both WT and FoxM1 CKO lungs exhibited similar increases in myeloperoxidase activity 6 hours after LPS challenge, indicating no difference in lung neutrophil sequestration. To determine the in vivo consequences of inhibiting post-LPS FoxM1 expression on vascular permeability, we isolated lungs from mice to quantify endothelial barrier function by measuring K_f,c and edema formation in isogravimetric lung preparation (24, 25). Mice were injected i.p. with either PBS (0.2 ml, pH 7.4) or LPS (5 mg/kg BW in 0.2 ml PBS), and lungs were evaluated at 6, 48, and 96 hours thereafter. We have previously shown that LPS at this concentration induces maximal increase of K_f,c at 6 hours in WT lungs (24). In the present study, we observed that K_f,c values at 6 hours were not significantly different in FoxM1 CKO and WT lungs (Figure 2A). In WT lungs, K_f,c returned to baseline level at 48 hours; however, K_f,c in FoxM1 CKO lungs remained significantly elevated at 48 hours and 96 hours. In addition, we recorded lung wet weight changes after LPS challenge as a measure of edema formation. As demonstrated by the precipitous rise in lung wet weight, edema formation was a prominent feature of FoxM1 CKO lungs beginning at 48 hours and continuing through 96 hours, a response not seen in WT lungs (Figure 2, B–D). Thus, LPS induced similar vascular injury in FoxM1 CKO and WT lungs as shown by the increase in K_f,c at 6 hours, but the critical difference seen in FoxM1 CKO lungs was the persistent vascular injury resulting in increased vascular permeability. We also observed perivascular leukocyte sequestration (primarily around large vessels) in FoxM1 CKO lungs 96 hours after LPS challenge; however, in WT lungs, this response was almost fully resolved by 96 hours (Figure 2, E and F). Strikingly, a higher dosage of LPS (12 mg/kg BW) resulted in a significantly greater mortality rate in FoxM1 CKO compared with WT mice (Figure 5A). We observed that 70% of FoxM1 CKO died within 96 hours of LPS treatment; in contrast, only 10% of WT mice died within this time (Figure 5A). Severe fulminant pulmonary edema was an important determinant of death in FoxM1 CKO mice (Figure 5B).

Figure 4

FoxM1 CKO mice exhibit endothelial apoptosis and lung neutrophil sequestration similar to that of WT mice in the early period following LPS challenge. (A) Representative micrographs of TUNEL staining. Cryosections of lungs (3- to 5-μm thick) collected 24 hours after LPS challenge were stained with FITC-conjugated TUNEL to identify apoptotic cells and anti-vWF antibody to identify endothelial cells; nuclei were counterstained with DAPI. Arrows, TUNEL⁺vWF⁺ cells. Scale bar: 25 mm. (B) Graphic representation of apoptotic endothelial cells and nonendothelial cells in FoxM1 CKO and WT lungs 24 hours after LPS challenge (5 mg/kg BW). The number of TUNEL-positive nuclei and vWF-positive cells from 3 to 4 consecutive cryosections from each mouse lung were averaged. Data are expressed as mean ± SD, n = 3–4 mice per group. There is no difference between WT and FoxM1 CKO. (C) Lung tissue myeloperoxidase (MPO) activity during basal period and 6 hours after LPS challenge (5 mg/kg BW). Data are expressed as mean ± SD, n = 4 mice per group. *P > 0.1 versus WT. Although there is a significant increase in MPO activity in mouse lungs 6 hours following LPS challenge, both WT and FoxM1 CKO lungs exhibit similar MPO activities, indicating a similar extent of polymorphonuclear leukocyte sequestration.

Figure 5

Increased mortality in FoxM1 CKO mice following LPS challenge. (A) Survival rate following LPS challenge. Mice were challenged i.p. with 12 mg/kg BW of LPS and housed under normal conditions. Approximately 70% of the FoxM1 CKO mice died from day 2 to day 4 after LPS challenge whereas only about 10% of WT mice died in this same period. Differences in the survival rates between the WT and CKO groups after LPS challenge were significant by Peto-Peto-Wilcoxon test. n = 11 mice per group. *P < 0.001. (B) Graphic representation of development of lung edema in FoxM1 CKO mice following LPS challenge (12 mg/kg). Lungs were collected either at 48 hours after LPS challenge (WT and CKO) or in some mice at day 2 just prior to their deaths (CKO-D). Data are expressed as mean ± SD. n = 3–5, ^#P < 0.05 versus WT; ^##P < 0.01 versus WT.

Deletion of FoxM1 in bone marrow cells and generation of proinflammatory cytokines and chemokines do not contribute to defective vascular repair in FoxM1 CKO mice. Tie2 promoter/enhancer-driven Cre expression may also induce the deletion of the FoxM1 gene in hematopoietic cells, which could thus interfere with the vascular repair process (31). As shown in Figure 6, FoxM1 expression was not different in FoxM1 CKO bone marrow cells compared with WT. However, it cannot be ruled out that FoxM1 might be deleted in a small subpopulation of hematopoietic cells in FoxM1 CKO. To determine whether FoxM1 CKO hematopoietic cells contributed to defective endothelial barrier repair following LPS challenge, WT bone marrow cells were transplanted into lethally irradiated FoxM1 CKO mice. FACS analysis demonstrated approximately 80% reconstitution of FoxM1 CKO bone marrow at 3 weeks after transplantation (Figure 6A). Transplantation of WT bone marrow in FoxM1 CKO mice failed to prevent the sustained increase of lung vascular permeability after LPS challenge (Figure 6B); thus, deletion of FoxM1 in hematopoietic cells does not contribute to defect in endothelial repair in FoxM1 CKO lungs.

Figure 6

Failure of WT bone marrow cells to rescue the vascular repair defect in FoxM1 CKO mice following LPS challenge. (A) FACS analysis of GFP-positive bone marrow lymphocytes (lymph) and total white blood cells. At 3 weeks following transplantation of WT bone marrow cells isolated from GFP-transgenic mice, bone marrow cells were isolated from reconstituted FoxM1 CKO and red blood cells were lysed prior to FACS analysis. Bone marrow white blood cells isolated from C57B6 mice and GFP-transgenic mice were used as negative and positive controls, respectively. Approximately 80% of bone marrow reconstitution was achieved. BMT, WT bone marrow cell–transplanted FoxM1 CKO. (B) Sustained increase in lung microvessel permeability in WT bone marrow cell–transplanted FoxM1 CKO following LPS challenge. At 5 weeks following WT bone marrow transplantation, lungs from WT or FoxM1 CKO mice were isolated for K_f,c measurement. Data are expressed as mean ± SD, n = 3–5 per group. *P < 0.001 versus basal; **P < 0.001, CKO versus WT. Basal, 0.2 ml PBS. CKO, FoxM1 CKO mice reconstituted with WT bone marrow cells. (C) Quantitative analysis of FoxM1 mRNA expression in bone marrow cells. Bone marrow white blood cells were isolated from either WT or FoxM1 CKO at the indicated time points following LPS challenge, and RNA was isolated for QRT-PCR analysis. FoxM1 mRNA levels were normalized to cyclophilin. Data are expressed as mean ± SD, n = 3 per group. In contrast to endothelial cells isolated from FoxM1 CKO lungs, the white blood cells from FoxM1 CKO bone marrow expressed FoxM1 at the same level as WT.

As increased expression of proinflammatory mediators may also contribute to increased vascular permeability in FoxM1 CKO lungs, we investigated this possibility by determining the expression of several mediators in lungs following LPS challenge. We observed no difference in expression of various cytokines and chemokines between WT and FoxM1 CKO lungs after LPS challenge (Table 1).

Table 1

Expression profiles of proinflammatory cytokines and chemokines in lungs following LPS challenge

Deletion of FoxM1in endothelial cells decreases cell proliferation after injury. We determined whether defective vascular repair in FoxM1 CKO mice was attributed to decreased endothelial cell proliferation. Mice were challenged with LPS, and cell proliferation was measured by BrdU incorporation. WT lungs exhibited increased cellular proliferation that peaked at 72 hours (Figure 7, A and B), paralleling the time course of FoxM1 mRNA induction (Figure 3A). Although cellular proliferation in FoxM1 CKO and WT lungs was comparable at 24 hours following LPS challenge, the response was significantly impaired in FoxM1 CKO lungs at 48 and 72 hours (Figure 7B). This impairment was primarily the result of the defect in endothelial cell proliferation (Figure 7C and Supplemental Figure 2). In tissue from FoxM1 CKO lungs, nuclear staining of p27^Kip1 protein, an inhibitor of cyclin-dependent kinases, was markedly increased in endothelial cells compared with WT endothelial cells after LPS challenge (Figure 8, A and B). Expression of p27^Kip1 was sustained in the FoxM1 CKO lungs while its expression decreased in WT lungs in response to LPS (Figure 8C and Supplemental Figure 3). QRT-PCR also demonstrated a significant decrease in the expression of Cdc25C and cyclin B1 in FoxM1 CKO lungs compared with WT following LPS challenge (Figure 8, D and E).

Figure 7

Decreased cell proliferation in FoxM1 CKO lungs following LPS-induced lung microvascular injury. (A) Representative micrographs of BrdU immunostaining. Cryosections of lungs (3- to 5-μm thick), collected 72 hours after LPS challenge, were stained with FITC-conjugated anti-BrdU antibody to identify proliferating cells; nuclei were counterstained with DAPI. Scale bar: 25 mm. (B) Graphic representation of decreased BrdU incorporation in FoxM1 CKO lungs and WT lungs at the indicated times following LPS challenge (5 mg/kg BW). From 3 to 4 consecutive cryosections from each mouse lung were examined, and the average number of BrdU-positive nuclei per 2,000 nuclei was used as the value for the mouse. Data are expressed as mean ± SD, n = 3–4 mice per time point. *P < 0.05 versus WT at 48 and 72 hours after LPS. There are few BrdU-positive nuclei (<1 BrdU positive nuclei per 2,000 nuclei) in both WT and FoxM1 CKO lungs in the absence of LPS challenge. (C) Graphic representation of decreased proliferating endothelial cells (double-positive staining of vWF and BrdU) in FoxM1 CKO lungs. Data are expressed as mean ± SD, percentage of total nuclei, n = 3–4 mice per time point per group. *P < 0.05 versus WT.

Figure 8

Increased nuclear staining and expression of p27^Kip1 and decreased expression of Cdc25C and cyclin B1 in FoxM1 CKO lungs. (A) Representative micrographs of immunostaining of p27^Kip1 and vWF. Cryosections of lungs collected 24 hours after LPS challenge (5 mg/kg) were stained with antibodies against p27^Kip1 and vWF (marker for endothelial cells); nuclei were counterstained with DAPI. Arrows indicate p27^Kip1+vWF⁺ cells. Scale bar: 25 μm. (B) Graphic representation of p27^Kip1-positive endothelial cells and nonendothelial cells. Data are expressed as mean ± SD, n = 3–4 mice per group. *P < 0.05 versus WT. (C) Sustained expression of p27^Kip1 in FoxM1 CKO lungs following LPS challenge (5 mg/kg). Protein (50 μg) from mouse lungs was loaded for each lane, the gel was electrophoresed, and proteins were transferred to PVDF and probed with monoclonal antibody against p27^Kip1. The same membrane was reprobed with a rabbit polyclonal antibody against β-actin as a loading control. The experiment was repeated 3 times with similar results. (D and E) Time course of expression of Cdc25C (D) and cyclin B1 (E) in lungs following LPS challenge (5 mg/kg). Total RNA was extracted from lungs collected from either WT mice or FoxM1 CKO mice after LPS exposure (5 mg/kg BW) at the indicated times. Quantitative analysis of mRNA levels in mouse lungs was performed with QRT-PCR. Cdc25C and cyclin B1 mRNA levels were normalized to cyclophilin. Data are expressed as mean ± SD, n = 3–4. *P < 0.05 versus WT.

FoxM1 is required for endothelial cell proliferation. The reduced endothelial cell proliferation found in FoxM1 CKO lungs suggests at FoxM1 plays a crucial role in vascular endothelial barrier reannealing. We addressed this question using endothelial cells isolated from either FoxM1 CKO or WT mouse lungs. The growth of endothelial cells cultured from FoxM1 CKO lungs was markedly reduced (Figure 9A). After 5–7 days in culture, FoxM1 CKO endothelial cells were less than 20% confluent whereas WT cells reached more than 80% confluency. In contrast, primary cultures of nonendothelial cells from FoxM1 CKO lungs, which retained the FoxM1^fl/fl targeted allele, displayed similar growth rates compared with WT (data not shown).

Figure 9

Defective cell cycle progression in FoxM1-deficient endothelial cells. (A) Representative micrographs of primary cultures of lung endothelial cells isolated from either 2-week-old WT or FoxM1 CKO mice. Phase contrast pictures were taken at day 6 following isolation and culture. Scale bars: 50 μm. Growth of FoxM1 CKO endothelial cells was severely reduced, and cell death was noted after 3–5 days in culture compared with WT cultures. (B) QRT-PCR demonstrated an approximately 80% decrease in levels of FoxM1 mRNA in HMEC-1 transfected with FoxM1 siRNA. Data are expressed as mean ± SD, n = 3. HMEC-1 were transfected with either 100 nM of siRNA (siRNA) against human FoxM1 or its cognate mutant siRNA (miRNA) or PBS (basal). At 65 hours after transfection, cells were collected for RNA isolation. (C) Cell-cycle profiles of HMEC-1. At 72 hours after transfection, cells were fixed and stained with propidium iodide. Representative FACS analyses of cell cycle profiles in asynchronous HMEC-1 are shown. (D) Graphic representation of accumulation of 4 N DNA content and increase of polyploid cells (8 N) in FoxM1-deficient endothelial cells. Data are expressed as mean ± SD, n = 4. *P < 0.01 siRNA versus either basal or miRNA; **P < 0.05 siRNA versus either basal or miRNA; ^ΧP < 0.05 siRNA versus either basal or miRNA.

To address mechanisms of the proliferative defect in FoxM1 CKO endothelial cells, we transfected siRNA into human microvascular endothelial cells 1 (HMEC-1) in order to suppress FoxM1 expression (Figure 9B). FACS analysis showed significant accumulation of 4 N DNA content in HMEC-1 cultures treated with FoxM1 siRNA (Figure 9, C and D), suggesting a defect in progression from G2 into mitosis. FoxM1-depleted HMEC-1 also exhibited a polyploid (8 N) peak indicative of endoreduplication caused by entry into a second round of DNA replication without completion of mitosis (Figure 9, C and D). As also seen in FoxM1 CKO lungs, the expression of p27^Kip1 was significantly upregulated and nuclear staining of p27^Kip1 was prominent in FoxM1-deficient endothelial cells (Figure 10, A and B). Both Cdk1 and Cdk2 activities were markedly decreased in FoxM1-deficient endothelial cells (Figure 10C), and the set of genes essential for cell cycle progression, including those coding for several cyclins, Cdc2, and Cdc25C, were significantly decreased (Figure 10D). Taken together, these results demonstrate that FoxM1 is required for endothelial cell proliferation and is therefore crucial for the recovery of endothelial barrier function after vascular injury.

Figure 10

FoxM1 regulates expression of a network of genes essential for cell cycle progression in endothelial cells. (A) Increased expression of the cell cycle inhibitor p27^Kip1 seen in endothelial cells following siRNA knockdown of FoxM1. HMEC-1 were transfected with either 100 nM siRNA against human FoxM1 (siRNA) or its cognate mutant siRNA (miRNA) or PBS (basal) for 72 hours, and protein lysates (50 μg/lane) were used for Western blotting. The experiment was repeated twice with similar results. (B) Representative micrographs of immunostaining of p27^Kip1 in human microvascular endothelial cells. FoxM1-deficient endothelial cells exhibited extensive nuclear staining of p27^Kip1. Scale bar: 25 μm. (C) Graphic representation of decreased activities of Cdk1 and Cdk2 in FoxM1-deficient human microvascular endothelial cells. At 65 hours after transfection, cells were lysed for the Cdk activity assay. Data are expressed as mean ± SD, n = 3. *P < 0.05 versus miRNA. (D) QRT-PCR demonstrated decreased expression of cyclins, Cdc2, and Cdc25C in FoxM1-deficient human microvascular endothelial cells. At 65 hours after transfection, total RNA was isolated for QRT-PCR analysis. Cyclophilin mRNA levels were used for normalization. Data are expressed as mean ± SD, n = 3. *P < 0.05 versus WT. Cyc, cyclin.

We have identified the essential role of FoxM1 in mediating endothelial repair following vascular injury using a mouse model of endothelial cell–restricted FoxM1 deletion. Mice with endothelial cell–specific deficiency of FoxM1 were viable and had normal basal endothelial barrier function. Lung microvascular permeability increased in these mice within 6 hours in response to the inflammatory mediator LPS. This response was similar to that of WT mice, as was LPS-induced endothelial cell apoptosis and neutrophil sequestration in the lungs. Strikingly, FoxM1 CKO exhibited a marked impairment in endothelial barrier repair. We also found a significant increase in mortality following LPS-induced injury in these mice. We attributed the persistent increase in lung microvascular permeability in the LPS-challenged FoxM1 CKO mice to defective endothelial cell proliferation that was associated with sustained expression of p27^Kip1 and decreased expression of cyclin B1 and Cdc25C. Together, these results demonstrate a crucial role of FoxM1 in restoring lung endothelial barrier integrity through its ability to activate endothelial cell proliferation in response to vascular endothelial injury.

There are several possible explanations for the prolonged increase of vascular permeability in FoxM1 CKO lungs following LPS challenge. It is possible that there may have been greater lung vascular injury in FoxM1 CKO mice; however, we observed similar increases in lung vascular permeability in CKO and WT mice within 6 hours of LPS challenge (time of peak increase in vascular permeability). The extent of endothelial cell apoptosis and neutrophil sequestration in lungs of WT and FoxM1 CKO mice was also the same; thus, differences in apoptosis and neutrophil infiltration are not likely causes of the prolonged increase in lung microvascular permeability seen in FoxM1 CKO mice. It is also possible that augmented production of proinflammatory mediators accounts for the defective endothelial repair in FoxM1 CKO mice. Our data also exclude this possibility, as FoxM1 CKO and WT lungs expressed these mediators (e.g., IL-1β, TNF-α, IFN-γ, macrophage inflammatory protein-1α [MIP-1α], and monocyte chemoattractant protein-1 [MCP-1]) to the same degree following LPS challenge. Based on the findings that Tie2 promoter/enhancer-driven Cre expression can cause gene deletion in endothelial cells as well as hematopoietic cells (31), it is possible that defective endothelial repair is the result of deletion of the FoxM1 gene in hematopoietic cells and failure of these cells to restore endothelial barrier function. However, we observed that FoxM1 was expressed in the bone marrow white blood cells of FoxM1 CKO mice at the same levels as in WT mice both basally and following LPS challenge. To rule out the possibility that FoxM1 expression was defective in a small subpopulation of hematopoietic cells of FoxM1 CKO mice, WT bone marrow cells were transplanted into lethally irradiated FoxM1 CKO mice. We found these reconstituted mutant mice still exhibited the sustained increase in vascular permeability at 48 hours after LPS challenge; thus, hematopoietic cells do not appear to be responsible for the defect seen in lung endothelial repair in FoxM1 CKO mice following vascular injury.

Our results support the hypothesis that endothelial injury induces the activation of a FoxM1-regulated program mediating endothelial cell proliferation and restoration of endothelial barrier integrity. We found FoxM1 mRNA was induced in lungs of WT mice at day 2 and the value peaked at day 3 following LPS challenge, that is, precisely during the recovery phase of LPS-induced vascular injury. The expression of FoxM1 was also correlated with the rate of cell proliferation in lungs following vascular injury induced by LPS. There was a significant defect in endothelial barrier reannealing in FoxM1 CKO lungs following LPS challenge, indicating that FoxM1 activation is critical for the restoration of endothelial integrity following lung vascular injury in mice.

To address whether cultured endothelial cells from FoxM1 CKO mice also have defective cell cycle progression, we isolated primary mouse endothelial cells from FoxM1 CKO lungs. These cells failed to proliferate normally, and many died after several days in culture. The suppression of FoxM1 expression in human microvascular endothelial cells using siRNA transfection also resulted in marked reduction in cell cycle progression. Our results documenting an important function of FoxM1 in restoration of endothelial integrity through its ability to promote cell proliferation are in agreement with the role of FoxM1 in regulating DNA replication and progression into and through mitosis (13, 14, 16, 17). Studies have demonstrated that FoxM1 regulates a network of genes essential for cell cycle progression in different cell types, including hepatocytes, mouse embryonic fibroblasts, and osteosarcoma U2OS (13, 14, 17–19). Our results are the first to establish the molecular basis of FoxM1-regulated proliferation of endothelial cells and vessel wall barrier repair.

We observed significant increases in protein levels of p27^Kip1 in FoxM1 CKO lungs following LPS challenge and in FoxM1-deficient human endothelial cells. p27^Kip1 prevents Cdk phosphorylation of retinoblastoma (Rb) required for the activation of the E2F transcription factors and thereby inhibits cell proliferation (32). Its expression is reduced in endothelial cells in response to mitogenic factor, e.g., vascular endothelial growth factor (33). In addition, we observed significant decreases in Cdk1 and Cdk2 activities in FoxM1-deficient endothelial cells due to inhibition in the expression of cyclin-dependent kinase Cdc2 and multiple cyclins, including cyclins F, A2, B1, and D2. Cyclins and Cdks are key regulators of the cell cycle (34). Different cyclin-Cdk complexes are involved in regulation of specific cell cycle transitions: cyclin D–Cdk4/6 for G1 progression, cyclin E–Cdk2 for G1/S transition, cyclin A–Cdk2 for S phase progression, and cyclins A/B–Cdk1 for entry into M phase (34). Progression through G2/M transition requires the activation of the Cdk1–cyclin B complex through dephosphorylation and activation of Cdk1 by M phase tyrosine phosphatases Cdc25B and Cdc25C (35). Cdc25C functions as a concentration-dependent inducer of mitotic control, and it directs dephosphorylation of cyclin B–bound Cdc2 and triggers entry into mitosis (35). In FoxM1-deficient endothelial cells, Cdc25C expression was markedly decreased. These data provide a mechanism for the defect in proliferation of endothelial cells seen in FoxM1 CKO mice. Thus, FoxM1 has an important role in regulating the expression of multiple genes essential for cell cycle progression in endothelial cells. In the absence of FoxM1, endothelial cells failed to proliferate, resulting in impaired recovery of endothelial barrier function.

In conclusion, the unperturbed vascular endothelium is normally quiescent and in a nonproliferative state. However, endothelial cells have the capacity to proliferate and reanneal the endothelial barrier after vascular injury. We provide evidence that the transcription factor FoxM1 is essential for the mechanism of endothelial cell proliferation following LPS-induced lung vascular injury and thereby plays a critical role in mediating endothelial barrier repair. Therefore, targeting FoxM1 activation to restore the restrictive nature of the vascular endothelial barrier may be a beneficial strategy for preventing lung inflammation associated with sepsis-induced ALI.

View Supplemental data

This work was supported by NIH grants T32HL007829, P01HL077806, and P01HL060678. We thank B. Kaushal and J.-S. Kim for their technical assistance.

Address correspondence to: Asrar B. Malik or You-Yang Zhao, Department of Pharmacology, University of Illinois College of Medicine, Chicago, Illinois 60612, USA. Phone: (312) 996-7635; Fax: (312) 996-1225; E-mail: abmalik@uic.edu (A.B. Malik). Phone: (312) 355-0238; Fax: (312) 996-1225; E-mail: yyzhao@uic.edu (Y.-Y. Zhao).

Nonstandard abbreviations used: ALI, acute lung injury; BW, body weight; Cdc25A, cell division cycle 25A; Cdk1, cyclin-dependent kinase 1; FoxM1, forkhead box M1; FoxM1 CKO, endothelial cell–restricted FoxM1-deficient (mice); HMEC-1, human microvascular endothelial cell(s) 1; K_f,c, filtration coefficient; MCP-1, monocyte chemoattractant protein-1; MIP-1α, macrophage inflammatory protein 1α; QRT-PCR, quantitative real-time RT-PCR; Rb, retinoblastoma.

Conflict of interest: The authors have declared that no conflict of interest exists.

Reference information: J. Clin. Invest.116:2333–2343 (2006). doi:10.1172/JCI27154.

See the related Commentary beginning on page 2316.

1. Cines, D.B., et al. 1998. Endothelial cells in physiology and in the pathophysiology of vascular disorders. Blood. 91:3527-3561.
   View this article via: PubMed Google Scholar
2. Ross, R. 1999. Atherosclerosis–an inflammatory disease. N. Engl. J. Med. 340:115-126.
   View this article via: PubMed CrossRef Google Scholar
3. Libby, P., Ridker, P.M., Maseri, A. 2002. Inflammation and atherosclerosis. Circulation. 105:1135-1143.
   View this article via: PubMed CrossRef Google Scholar
4. Austin, G.E., Ratliff, N.B., Hollman, J., Tabei, S., Phillips, D.F. 1985. Intimal proliferation of smooth muscle cells as an explanation for recurrent coronary artery stenosis after percutaneous transluminal coronary angioplasty. J. Am. Coll. Cardiol. 6:369-375.
   View this article via: PubMed Google Scholar
5. Schwartz, R.S., Holmes, D.R., Topol, E.J. 1992. The restenosis paradigm revisited: an alternative proposal for cellular mechanisms. J. Am. Coll. Cardiol. 20:1284-1293.
   View this article via: PubMed Google Scholar
6. McFadden, E.P., et al. 1993. Response of human coronary arteries to serotonin after injury by coronary angioplasty. Circulation. 88:2076-2085.
   View this article via: PubMed Google Scholar
7. Ware, L.B., Matthay, M.A. 2000. The acute respiratory distress syndrome. N. Engl. J. Med. 342:1334-1349.
   View this article via: PubMed CrossRef Google Scholar
8. Kaestner, K.H., Knochel, W., Martinez, D.E. 2000. Unified nomenclature for the winged helix/forkhead transcription factors. Genes Dev. 14:142-146.
   View this article via: PubMed Google Scholar
9. Clark, K.L., Halay, E.D., Lai, E., Burley, S.K. 1993. Co-crystal structure of the HNF-3/fork head DNA-recognition motif resembles histone H5. Nature. 364:412-420.
   View this article via: PubMed CrossRef Google Scholar
10. Korver, W., Roose, J., Clevers, H. 1997. The winged-helix transcription factor Trident is expressed in cycling cells. Nucleic Acids Res. 25:1715-1719.
    View this article via: PubMed CrossRef Google Scholar
11. Ye, H., et al. 1997. Hepatocyte nuclear factor 3/fork head homolog 11 is expressed in proliferating epithelial and mesenchymal cells of embryonic and adult tissues. Mol. Cell. Biol. 17:1626-1641.
    View this article via: PubMed Google Scholar
12. Yao, K.M., Sha, M., Lu, Z., Wong, G.G. 1997. Molecular analysis of a novel winged helix protein, WIN. Expression pattern, DNA binding property, and alternative splicing within the DNA binding domain. J. Biol. Chem. 272:19827-19836.
    View this article via: PubMed CrossRef Google Scholar
13. Wang, X., Kiyokawa, H., Dennewitz, M.B., Costa, R.H. 2002. The Forkhead Box m1b transcription factor is essential for hepatocyte DNA replication and mitosis during mouse liver regeneration. Proc. Natl. Acad. Sci. U. S. A. 99:16881-16886.
    View this article via: PubMed CrossRef Google Scholar
14. Kalinichenko, V.V., et al. 2004. Foxm1b transcription factor is essential for development of hepatocellular carcinomas and is negatively regulated by the p19ARF tumor suppressor. Genes Dev. 18:830-850.
    View this article via: PubMed CrossRef Google Scholar
15. Kalinichenko, V.V., et al. 2003. Ubiquitous expression of the forkhead box M1B transgene accelerates proliferation of distinct pulmonary cell types following lung injury. J. Biol. Chem. 278:37888-37894.
    View this article via: PubMed CrossRef Google Scholar
16. Krupczak-Hollis, K., et al. 2004. The mouse Forkhead Box m1 transcription factor is essential for hepatoblast mitosis and development of intrahepatic bile ducts and vessels during liver morphogenesis. Dev. Biol. 276:74-88.
    View this article via: PubMed CrossRef Google Scholar
17. Laoukili, J., et al. 2005. FoxM1 is required for execution of the mitotic programme and chromosome stability. Nat. Cell Biol. 7:126-136.
    View this article via: PubMed CrossRef Google Scholar
18. Costa, R.H. 2005. FoxM1 dances with mitosis. Nat. Cell Biol. 7:108-110.
    View this article via: PubMed CrossRef Google Scholar
19. Wang, I.-C., et al. 2005. Forkhead box M1 regulates the transcriptional network of genes essential for mitotic progression and genes encoding the SCF (Skp2-Cks1) ubiquitin ligase. Mol. Cell. Biol. 25:10875-10894.
    View this article via: PubMed CrossRef Google Scholar
20. Kisanuki, Y.Y., et al. 2001. Tie2-Cre transgenic mice: a new model for endothelial cell-lineage analysis in vivo. Dev. Biol. 230:230-242.
    View this article via: PubMed CrossRef Google Scholar
21. Liao, Y., Day, K.H., Damon, D.N., Duling, B.R. 2001. Endothelial cell-specific knockout of connexin 43 causes hypotension and bradycardia in mice. Proc. Natl. Acad. Sci. U. S. A. 98:9989-9994.
    View this article via: PubMed CrossRef Google Scholar
22. Kondo, T., et al. 2003. Knockout of insulin and IGF-1 receptors on vascular endothelial cells protects against retinal neovascularization. J. Clin. Invest. 111:1835-1842.
    View this article via: JCI CrossRef PubMed Google Scholar
23. Kim, I.M., et al. 2005. The forkhead box m1 transcription factor is essential for embryonic development of pulmonary vasculature. J. Biol. Chem. 280:22278-22286.
    View this article via: PubMed CrossRef Google Scholar
24. Gao, X., et al. 2001. Differential role of CD18 integrins in mediating lung neutrophil sequestration and increased microvascular permeability induced by Escherichia coli in mice. J. Immunol. 167:2895-2901.
    View this article via: PubMed Google Scholar
25. Tiruppathi, C., et al. 2002. Impairment of store-operated Ca2+ entry in TRPC4(–/–) mice interferes with increase in lung microvascular permeability. Circ. Res. 91:70-76.
    View this article via: PubMed CrossRef Google Scholar
26. Windsor, A.C., Mullen, P.G., Fowler, A.A. 1993. Acute lung injury: what have we learned from animal models? Am. J. Med. Sci. 306:111-116.
    View this article via: PubMed Google Scholar
27. Ware, L.B., Matthay, M.A. 2000. The acute respiratory distress syndrome. N. Engl. J. Med. 342:1334-1349.
    View this article via: PubMed CrossRef Google Scholar
28. Fujita, M., et al. 1998. Endothelial cell apoptosis in lipopolysaccharide-induced lung injury in mice. Int. Arch. Allergy Immunol. 117:202-208.
    View this article via: PubMed CrossRef Google Scholar
29. Kitamura, Y., et al. 2001. Fas/FasL-dependent apoptosis of alveolar cells after lipopolysaccharide-induced lung injury in mice. Am. J. Respir. Crit. Care Med. 163:762-769.
    View this article via: PubMed Google Scholar
30. Wiener-Kronish, J.P., Albertine, K.H., Matthay, M.A. 1991. Differential responses of the endothelial and epithelial barriers of the lung in sheep to Escherichia coli endotoxin. J. Clin. Invest. 88:864-875.
    View this article via: JCI PubMed CrossRef Google Scholar
31. Koni, P.A., et al. 2001. Conditional vascular cell adhesion molecule 1 deletion in mice: impaired lymphocyte migration to bone marrow. J. Exp. Med. 193:741-753.
    View this article via: PubMed CrossRef Google Scholar
32. Sherr, C.J., Roberts, J.M. 1999. CDK inhibitors: positive and negative regulators of G1-phase progression. Genes Dev. 13:1501-1512.
    View this article via: PubMed CrossRef Google Scholar
33. Abid, M.R., et al. 2004. Vascular endothelial growth factor activates PI3K/Akt/Forkhead signaling in endothelial cells. Arterioscler. Thromb. Vasc. Biol. 24:294-300.
    View this article via: PubMed CrossRef Google Scholar
34. Massague, J. 2004. G1 cell-cycle control and cancer. Nature. 432:298-306.
    View this article via: PubMed CrossRef Google Scholar
35. Nilsson, I., Hoffman, I. 2000. Cell cycle regulation by the Cdc25 phosphatase family. Prog. Cell Cycle Res. 4:107-114.
    View this article via: PubMed Google Scholar
36. Zhao, Y.Y., et al. 2002. Defects in caveolin-1 cause dilated cardiomyopathy and pulmonary hypertension in knockout mice. Proc. Natl. Acad. Sci. U. S. A. 99:11375-11380.
    View this article via: PubMed CrossRef Google Scholar
37. Frey, R.S., Rahman, A., Kefer, J.C., Minshall, R.D., Malik, A.B. 2002. PKCzeta regulates TNF-alpha-induced activation of NADPH oxidase in endothelial cells. Circ. Res. 90:1012-1019..
    View this article via: PubMed CrossRef Google Scholar